{ "best_metric": 0.31609994173049927, "best_model_checkpoint": "saves/LLaMA2-7B-Chat/lora/2023-08-25-11-47-37/checkpoint-500", "epoch": 0.8565310492505354, "global_step": 600, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.01, "learning_rate": 4.9999300626531746e-05, "logits/chosen": -0.7403622269630432, "logits/rejected": -0.6866486668586731, "logps/chosen": -181.39700317382812, "logps/rejected": -208.150146484375, "loss": 0.6892, "rewards/accuracies": 0.44999998807907104, "rewards/chosen": -5.543716179090552e-05, "rewards/margins": 0.008117685094475746, "rewards/rejected": -0.008173122070729733, "step": 5 }, { "epoch": 0.01, "learning_rate": 4.999720254525684e-05, "logits/chosen": -0.7327243089675903, "logits/rejected": -0.6666526794433594, "logps/chosen": -231.14956665039062, "logps/rejected": -259.27978515625, "loss": 0.6702, "rewards/accuracies": 0.6625000238418579, "rewards/chosen": -0.010688358917832375, "rewards/margins": 0.04831721633672714, "rewards/rejected": -0.05900556966662407, "step": 10 }, { "epoch": 0.02, "learning_rate": 4.9995272362322506e-05, "logits/chosen": -0.7042385935783386, "logits/rejected": -0.6349071264266968, "logps/chosen": -188.58367919921875, "logps/rejected": -223.201416015625, "loss": 0.6677, "rewards/accuracies": 0.625, "rewards/chosen": -0.07146739959716797, "rewards/margins": 0.05914110690355301, "rewards/rejected": -0.13060849905014038, "step": 15 }, { "epoch": 0.03, "learning_rate": 4.999093662527553e-05, "logits/chosen": -0.6996177434921265, "logits/rejected": -0.6077739000320435, "logps/chosen": -196.89700317382812, "logps/rejected": -243.52450561523438, "loss": 0.6027, "rewards/accuracies": 0.737500011920929, "rewards/chosen": -0.0204925499856472, "rewards/margins": 0.22103042900562286, "rewards/rejected": -0.24152295291423798, "step": 20 }, { "epoch": 0.04, "learning_rate": 4.998520264838675e-05, "logits/chosen": -0.6775780916213989, "logits/rejected": -0.6084710359573364, "logps/chosen": -194.9524688720703, "logps/rejected": -232.9259490966797, "loss": 0.584, "rewards/accuracies": 0.762499988079071, "rewards/chosen": -0.22763672471046448, "rewards/margins": 0.2767763137817383, "rewards/rejected": -0.5044130086898804, "step": 25 }, { "epoch": 0.04, "learning_rate": 4.997807075247146e-05, "logits/chosen": -0.6930967569351196, "logits/rejected": -0.624418318271637, "logps/chosen": -201.62631225585938, "logps/rejected": -231.8871307373047, "loss": 0.5054, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -0.42100492119789124, "rewards/margins": 0.5007229447364807, "rewards/rejected": -0.9217279553413391, "step": 30 }, { "epoch": 0.05, "learning_rate": 4.996954133655838e-05, "logits/chosen": -0.6575398445129395, "logits/rejected": -0.5734818577766418, "logps/chosen": -215.608154296875, "logps/rejected": -259.5364990234375, "loss": 0.491, "rewards/accuracies": 0.737500011920929, "rewards/chosen": -1.129244089126587, "rewards/margins": 0.6850830316543579, "rewards/rejected": -1.8143272399902344, "step": 35 }, { "epoch": 0.06, "learning_rate": 4.9959614877867264e-05, "logits/chosen": -0.5865448713302612, "logits/rejected": -0.5297074317932129, "logps/chosen": -251.30545043945312, "logps/rejected": -282.8074035644531, "loss": 0.4294, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -1.6625725030899048, "rewards/margins": 0.9844868779182434, "rewards/rejected": -2.647059440612793, "step": 40 }, { "epoch": 0.06, "learning_rate": 4.994829193178228e-05, "logits/chosen": -0.6305592656135559, "logits/rejected": -0.5662962794303894, "logps/chosen": -238.34683227539062, "logps/rejected": -288.0771789550781, "loss": 0.4195, "rewards/accuracies": 0.8125, "rewards/chosen": -2.0896081924438477, "rewards/margins": 1.2067267894744873, "rewards/rejected": -3.296334743499756, "step": 45 }, { "epoch": 0.07, "learning_rate": 4.9935573131820854e-05, "logits/chosen": -0.5488190054893494, "logits/rejected": -0.4458453059196472, "logps/chosen": -237.5546417236328, "logps/rejected": -296.1310119628906, "loss": 0.2968, "rewards/accuracies": 0.925000011920929, "rewards/chosen": -2.8281638622283936, "rewards/margins": 1.8660491704940796, "rewards/rejected": -4.694213390350342, "step": 50 }, { "epoch": 0.08, "learning_rate": 4.99214591895983e-05, "logits/chosen": -0.496940940618515, "logits/rejected": -0.39123407006263733, "logps/chosen": -235.4664764404297, "logps/rejected": -301.9604797363281, "loss": 0.3675, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -4.971009731292725, "rewards/margins": 2.1594393253326416, "rewards/rejected": -7.130448818206787, "step": 55 }, { "epoch": 0.09, "learning_rate": 4.990595089478795e-05, "logits/chosen": -0.44920119643211365, "logits/rejected": -0.3571397066116333, "logps/chosen": -270.9496154785156, "logps/rejected": -346.5156555175781, "loss": 0.3613, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -5.407833576202393, "rewards/margins": 2.685551166534424, "rewards/rejected": -8.093385696411133, "step": 60 }, { "epoch": 0.09, "learning_rate": 4.9889049115077005e-05, "logits/chosen": -0.5083180665969849, "logits/rejected": -0.4510710835456848, "logps/chosen": -266.8662109375, "logps/rejected": -305.4253845214844, "loss": 0.4624, "rewards/accuracies": 0.8125, "rewards/chosen": -5.9966888427734375, "rewards/margins": 2.306278705596924, "rewards/rejected": -8.302966117858887, "step": 65 }, { "epoch": 0.1, "learning_rate": 4.987075479611796e-05, "logits/chosen": -0.45440906286239624, "logits/rejected": -0.38472697138786316, "logps/chosen": -277.6520690917969, "logps/rejected": -340.0121765136719, "loss": 0.4809, "rewards/accuracies": 0.762499988079071, "rewards/chosen": -6.467832088470459, "rewards/margins": 2.306380271911621, "rewards/rejected": -8.774212837219238, "step": 70 }, { "epoch": 0.11, "learning_rate": 4.9851068961475725e-05, "logits/chosen": -0.44587111473083496, "logits/rejected": -0.3460499048233032, "logps/chosen": -248.66030883789062, "logps/rejected": -324.4167785644531, "loss": 0.3326, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -5.31089973449707, "rewards/margins": 2.787933826446533, "rewards/rejected": -8.098833084106445, "step": 75 }, { "epoch": 0.11, "learning_rate": 4.982999271257033e-05, "logits/chosen": -0.47116953134536743, "logits/rejected": -0.40097618103027344, "logps/chosen": -300.4339904785156, "logps/rejected": -352.66363525390625, "loss": 0.3232, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -5.735043525695801, "rewards/margins": 2.7127461433410645, "rewards/rejected": -8.447790145874023, "step": 80 }, { "epoch": 0.12, "learning_rate": 4.980752722861531e-05, "logits/chosen": -0.5418936014175415, "logits/rejected": -0.48130369186401367, "logps/chosen": -273.4432067871094, "logps/rejected": -305.4323425292969, "loss": 0.4748, "rewards/accuracies": 0.7749999761581421, "rewards/chosen": -6.1163129806518555, "rewards/margins": 1.9569565057754517, "rewards/rejected": -8.07326889038086, "step": 85 }, { "epoch": 0.13, "learning_rate": 4.978367376655177e-05, "logits/chosen": -0.6019693613052368, "logits/rejected": -0.501119077205658, "logps/chosen": -241.8748016357422, "logps/rejected": -303.2873840332031, "loss": 0.3463, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -4.80649471282959, "rewards/margins": 2.6070265769958496, "rewards/rejected": -7.413522243499756, "step": 90 }, { "epoch": 0.14, "learning_rate": 4.9758433660977964e-05, "logits/chosen": -0.5586158633232117, "logits/rejected": -0.46125540137290955, "logps/chosen": -253.9844512939453, "logps/rejected": -318.15960693359375, "loss": 0.2994, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -4.774440288543701, "rewards/margins": 2.6352882385253906, "rewards/rejected": -7.40972900390625, "step": 95 }, { "epoch": 0.14, "learning_rate": 4.9731808324074717e-05, "logits/chosen": -0.5857738256454468, "logits/rejected": -0.5078204274177551, "logps/chosen": -244.3525390625, "logps/rejected": -299.4289855957031, "loss": 0.4066, "rewards/accuracies": 0.800000011920929, "rewards/chosen": -4.955862998962402, "rewards/margins": 2.2310631275177, "rewards/rejected": -7.186926364898682, "step": 100 }, { "epoch": 0.14, "eval_logits/chosen": -0.6736606359481812, "eval_logits/rejected": -0.5934695601463318, "eval_logps/chosen": -269.2135009765625, "eval_logps/rejected": -330.87744140625, "eval_loss": 0.36347857117652893, "eval_rewards/accuracies": 0.8051801919937134, "eval_rewards/chosen": -5.528914451599121, "eval_rewards/margins": 2.5316162109375, "eval_rewards/rejected": -8.060530662536621, "eval_runtime": 618.2599, "eval_samples_per_second": 0.954, "eval_steps_per_second": 0.12, "step": 100 }, { "epoch": 0.15, "learning_rate": 4.970379924552635e-05, "logits/chosen": -0.6208111047744751, "logits/rejected": -0.5168194770812988, "logps/chosen": -256.57330322265625, "logps/rejected": -336.573486328125, "loss": 0.2796, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -5.576841354370117, "rewards/margins": 2.8258020877838135, "rewards/rejected": -8.402643203735352, "step": 105 }, { "epoch": 0.16, "learning_rate": 4.9674407992437394e-05, "logits/chosen": -0.5667734742164612, "logits/rejected": -0.5014970302581787, "logps/chosen": -279.5911560058594, "logps/rejected": -323.6059875488281, "loss": 0.3282, "rewards/accuracies": 0.875, "rewards/chosen": -6.653090476989746, "rewards/margins": 2.4938788414001465, "rewards/rejected": -9.14696979522705, "step": 110 }, { "epoch": 0.16, "learning_rate": 4.964363620924485e-05, "logits/chosen": -0.5388852953910828, "logits/rejected": -0.4672119617462158, "logps/chosen": -275.47357177734375, "logps/rejected": -326.81182861328125, "loss": 0.4124, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -6.149296760559082, "rewards/margins": 2.311483860015869, "rewards/rejected": -8.460780143737793, "step": 115 }, { "epoch": 0.17, "learning_rate": 4.961148561762622e-05, "logits/chosen": -0.5033109784126282, "logits/rejected": -0.405276358127594, "logps/chosen": -256.9668884277344, "logps/rejected": -325.78839111328125, "loss": 0.3181, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -6.381417274475098, "rewards/margins": 2.8026371002197266, "rewards/rejected": -9.184053421020508, "step": 120 }, { "epoch": 0.18, "learning_rate": 4.9577958016403156e-05, "logits/chosen": -0.5019578337669373, "logits/rejected": -0.42507410049438477, "logps/chosen": -298.7206115722656, "logps/rejected": -352.7537841796875, "loss": 0.4869, "rewards/accuracies": 0.7124999761581421, "rewards/chosen": -6.763588905334473, "rewards/margins": 2.4632394313812256, "rewards/rejected": -9.226827621459961, "step": 125 }, { "epoch": 0.19, "learning_rate": 4.954305528144085e-05, "logits/chosen": -0.5088043808937073, "logits/rejected": -0.40408769249916077, "logps/chosen": -260.7892761230469, "logps/rejected": -330.84747314453125, "loss": 0.2442, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -6.382019996643066, "rewards/margins": 3.018307685852051, "rewards/rejected": -9.400327682495117, "step": 130 }, { "epoch": 0.19, "learning_rate": 4.9506779365543046e-05, "logits/chosen": -0.4910973608493805, "logits/rejected": -0.4001993238925934, "logps/chosen": -273.5351867675781, "logps/rejected": -341.2701110839844, "loss": 0.3465, "rewards/accuracies": 0.8125, "rewards/chosen": -6.231161117553711, "rewards/margins": 2.6937344074249268, "rewards/rejected": -8.924895286560059, "step": 135 }, { "epoch": 0.2, "learning_rate": 4.946913229834279e-05, "logits/chosen": -0.5775797367095947, "logits/rejected": -0.48806411027908325, "logps/chosen": -259.29351806640625, "logps/rejected": -339.4101257324219, "loss": 0.3707, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": -6.118729114532471, "rewards/margins": 3.231736421585083, "rewards/rejected": -9.350464820861816, "step": 140 }, { "epoch": 0.21, "learning_rate": 4.943011618618887e-05, "logits/chosen": -0.4882396161556244, "logits/rejected": -0.4096860885620117, "logps/chosen": -258.9165954589844, "logps/rejected": -323.74822998046875, "loss": 0.3093, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -5.742823600769043, "rewards/margins": 3.2949295043945312, "rewards/rejected": -9.037752151489258, "step": 145 }, { "epoch": 0.21, "learning_rate": 4.938973321202799e-05, "logits/chosen": -0.5276007056236267, "logits/rejected": -0.4654686450958252, "logps/chosen": -320.4021301269531, "logps/rejected": -362.60992431640625, "loss": 0.4327, "rewards/accuracies": 0.762499988079071, "rewards/chosen": -7.360659122467041, "rewards/margins": 2.3584816455841064, "rewards/rejected": -9.719141006469727, "step": 150 }, { "epoch": 0.22, "learning_rate": 4.93479856352826e-05, "logits/chosen": -0.6080215573310852, "logits/rejected": -0.561368465423584, "logps/chosen": -260.1334533691406, "logps/rejected": -304.76434326171875, "loss": 0.4775, "rewards/accuracies": 0.7749999761581421, "rewards/chosen": -5.641587257385254, "rewards/margins": 2.074352264404297, "rewards/rejected": -7.715939521789551, "step": 155 }, { "epoch": 0.23, "learning_rate": 4.930487579172451e-05, "logits/chosen": -0.6534048318862915, "logits/rejected": -0.557198166847229, "logps/chosen": -238.5316925048828, "logps/rejected": -303.6593017578125, "loss": 0.3279, "rewards/accuracies": 0.8125, "rewards/chosen": -4.701598167419434, "rewards/margins": 2.6716597080230713, "rewards/rejected": -7.373257637023926, "step": 160 }, { "epoch": 0.24, "learning_rate": 4.926040609334418e-05, "logits/chosen": -0.5949841141700745, "logits/rejected": -0.4924190938472748, "logps/chosen": -268.88604736328125, "logps/rejected": -336.0675048828125, "loss": 0.3907, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -5.271938800811768, "rewards/margins": 2.7763962745666504, "rewards/rejected": -8.048335075378418, "step": 165 }, { "epoch": 0.24, "learning_rate": 4.9214579028215776e-05, "logits/chosen": -0.5787319540977478, "logits/rejected": -0.5097864270210266, "logps/chosen": -270.76904296875, "logps/rejected": -330.5365905761719, "loss": 0.4062, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -5.298426628112793, "rewards/margins": 2.6809771060943604, "rewards/rejected": -7.979403495788574, "step": 170 }, { "epoch": 0.25, "learning_rate": 4.9167397160358e-05, "logits/chosen": -0.5979365110397339, "logits/rejected": -0.5267354846000671, "logps/chosen": -250.3006591796875, "logps/rejected": -310.52764892578125, "loss": 0.3127, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -5.677347660064697, "rewards/margins": 2.7255032062530518, "rewards/rejected": -8.402850151062012, "step": 175 }, { "epoch": 0.26, "learning_rate": 4.911886312959055e-05, "logits/chosen": -0.600624144077301, "logits/rejected": -0.5323041677474976, "logps/chosen": -263.458984375, "logps/rejected": -324.83404541015625, "loss": 0.2828, "rewards/accuracies": 0.875, "rewards/chosen": -5.505527496337891, "rewards/margins": 2.577638626098633, "rewards/rejected": -8.083166122436523, "step": 180 }, { "epoch": 0.26, "learning_rate": 4.906897965138653e-05, "logits/chosen": -0.6335456371307373, "logits/rejected": -0.5383679270744324, "logps/chosen": -271.63812255859375, "logps/rejected": -345.01702880859375, "loss": 0.3611, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -6.672010898590088, "rewards/margins": 3.4995288848876953, "rewards/rejected": -10.171539306640625, "step": 185 }, { "epoch": 0.27, "learning_rate": 4.90177495167204e-05, "logits/chosen": -0.6199727058410645, "logits/rejected": -0.5486131906509399, "logps/chosen": -282.41583251953125, "logps/rejected": -347.9653625488281, "loss": 0.4037, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -6.799245357513428, "rewards/margins": 2.8858962059020996, "rewards/rejected": -9.685141563415527, "step": 190 }, { "epoch": 0.28, "learning_rate": 4.896517559191194e-05, "logits/chosen": -0.683984100818634, "logits/rejected": -0.5775423049926758, "logps/chosen": -262.58538818359375, "logps/rejected": -335.623779296875, "loss": 0.4663, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -7.013031959533691, "rewards/margins": 3.129578113555908, "rewards/rejected": -10.142609596252441, "step": 195 }, { "epoch": 0.29, "learning_rate": 4.891126081846579e-05, "logits/chosen": -0.6567696332931519, "logits/rejected": -0.5673569440841675, "logps/chosen": -256.83892822265625, "logps/rejected": -325.61065673828125, "loss": 0.3115, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -6.3370866775512695, "rewards/margins": 3.1089577674865723, "rewards/rejected": -9.446043968200684, "step": 200 }, { "epoch": 0.29, "eval_logits/chosen": -0.7238260507583618, "eval_logits/rejected": -0.6414477229118347, "eval_logps/chosen": -288.9779968261719, "eval_logps/rejected": -360.12066650390625, "eval_loss": 0.3467591404914856, "eval_rewards/accuracies": 0.832207202911377, "eval_rewards/chosen": -7.505363464355469, "eval_rewards/margins": 3.479483127593994, "eval_rewards/rejected": -10.984847068786621, "eval_runtime": 617.3232, "eval_samples_per_second": 0.956, "eval_steps_per_second": 0.12, "step": 200 }, { "epoch": 0.29, "learning_rate": 4.8856008212906925e-05, "logits/chosen": -0.7143681049346924, "logits/rejected": -0.6183390021324158, "logps/chosen": -269.6623229980469, "logps/rejected": -350.52459716796875, "loss": 0.2657, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -7.174610137939453, "rewards/margins": 3.6835849285125732, "rewards/rejected": -10.858195304870605, "step": 205 }, { "epoch": 0.3, "learning_rate": 4.879942086661184e-05, "logits/chosen": -0.6396993398666382, "logits/rejected": -0.5612179040908813, "logps/chosen": -320.8450622558594, "logps/rejected": -394.2232360839844, "loss": 0.2258, "rewards/accuracies": 0.9125000238418579, "rewards/chosen": -8.637298583984375, "rewards/margins": 3.9497833251953125, "rewards/rejected": -12.587082862854004, "step": 210 }, { "epoch": 0.31, "learning_rate": 4.8741501945635656e-05, "logits/chosen": -0.6102300882339478, "logits/rejected": -0.5374587178230286, "logps/chosen": -261.81256103515625, "logps/rejected": -329.33453369140625, "loss": 0.3376, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -7.931007385253906, "rewards/margins": 3.3228728771209717, "rewards/rejected": -11.253880500793457, "step": 215 }, { "epoch": 0.31, "learning_rate": 4.8682254690534876e-05, "logits/chosen": -0.589969277381897, "logits/rejected": -0.5134058594703674, "logps/chosen": -284.9112243652344, "logps/rejected": -351.90643310546875, "loss": 0.3639, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -7.687941551208496, "rewards/margins": 3.43585205078125, "rewards/rejected": -11.123793601989746, "step": 220 }, { "epoch": 0.32, "learning_rate": 4.862168241618617e-05, "logits/chosen": -0.5980569124221802, "logits/rejected": -0.5199486017227173, "logps/chosen": -243.0383758544922, "logps/rejected": -318.08319091796875, "loss": 0.371, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -6.680181980133057, "rewards/margins": 3.751661777496338, "rewards/rejected": -10.431844711303711, "step": 225 }, { "epoch": 0.33, "learning_rate": 4.8559788511600876e-05, "logits/chosen": -0.5468825101852417, "logits/rejected": -0.46182718873023987, "logps/chosen": -300.0223388671875, "logps/rejected": -365.42559814453125, "loss": 0.3014, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -8.608762741088867, "rewards/margins": 3.9094433784484863, "rewards/rejected": -12.518205642700195, "step": 230 }, { "epoch": 0.34, "learning_rate": 4.849657643973535e-05, "logits/chosen": -0.5835626721382141, "logits/rejected": -0.46411094069480896, "logps/chosen": -289.1901550292969, "logps/rejected": -391.5763244628906, "loss": 0.2262, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -7.415358066558838, "rewards/margins": 4.35217809677124, "rewards/rejected": -11.767536163330078, "step": 235 }, { "epoch": 0.34, "learning_rate": 4.843204973729729e-05, "logits/chosen": -0.5552169680595398, "logits/rejected": -0.4667941927909851, "logps/chosen": -262.4269714355469, "logps/rejected": -344.49462890625, "loss": 0.318, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -6.991745948791504, "rewards/margins": 3.809674024581909, "rewards/rejected": -10.801420211791992, "step": 240 }, { "epoch": 0.35, "learning_rate": 4.8366212014547775e-05, "logits/chosen": -0.5869508981704712, "logits/rejected": -0.536375880241394, "logps/chosen": -263.1332092285156, "logps/rejected": -323.9794006347656, "loss": 0.3479, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -6.706090450286865, "rewards/margins": 3.0286171436309814, "rewards/rejected": -9.734708786010742, "step": 245 }, { "epoch": 0.36, "learning_rate": 4.8299066955099335e-05, "logits/chosen": -0.6492255330085754, "logits/rejected": -0.572007954120636, "logps/chosen": -283.6471862792969, "logps/rejected": -338.55535888671875, "loss": 0.5536, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": -6.6781325340271, "rewards/margins": 1.9866206645965576, "rewards/rejected": -8.664752960205078, "step": 250 }, { "epoch": 0.36, "learning_rate": 4.823061831570981e-05, "logits/chosen": -0.5709847807884216, "logits/rejected": -0.49480828642845154, "logps/chosen": -280.56005859375, "logps/rejected": -343.18511962890625, "loss": 0.3881, "rewards/accuracies": 0.800000011920929, "rewards/chosen": -6.904370307922363, "rewards/margins": 2.5633044242858887, "rewards/rejected": -9.46767520904541, "step": 255 }, { "epoch": 0.37, "learning_rate": 4.816086992607219e-05, "logits/chosen": -0.5884816646575928, "logits/rejected": -0.48144420981407166, "logps/chosen": -259.3785095214844, "logps/rejected": -330.3200988769531, "loss": 0.3031, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -6.253561973571777, "rewards/margins": 2.664248466491699, "rewards/rejected": -8.917810440063477, "step": 260 }, { "epoch": 0.38, "learning_rate": 4.8089825688600324e-05, "logits/chosen": -0.6093907356262207, "logits/rejected": -0.5078362226486206, "logps/chosen": -264.2926025390625, "logps/rejected": -339.7705993652344, "loss": 0.2332, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": -6.6783013343811035, "rewards/margins": 3.519818067550659, "rewards/rejected": -10.198118209838867, "step": 265 }, { "epoch": 0.39, "learning_rate": 4.8017489578210604e-05, "logits/chosen": -0.5917445421218872, "logits/rejected": -0.49844178557395935, "logps/chosen": -271.13677978515625, "logps/rejected": -333.9804382324219, "loss": 0.4124, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -7.525943756103516, "rewards/margins": 2.7923340797424316, "rewards/rejected": -10.318277359008789, "step": 270 }, { "epoch": 0.39, "learning_rate": 4.794386564209953e-05, "logits/chosen": -0.6364978551864624, "logits/rejected": -0.5267010927200317, "logps/chosen": -282.91021728515625, "logps/rejected": -378.1858825683594, "loss": 0.2063, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": -6.725545406341553, "rewards/margins": 4.508333206176758, "rewards/rejected": -11.233880043029785, "step": 275 }, { "epoch": 0.4, "learning_rate": 4.786895799951732e-05, "logits/chosen": -0.5923280119895935, "logits/rejected": -0.512219250202179, "logps/chosen": -285.5169372558594, "logps/rejected": -352.58148193359375, "loss": 0.3159, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -8.174338340759277, "rewards/margins": 4.347126007080078, "rewards/rejected": -12.521463394165039, "step": 280 }, { "epoch": 0.41, "learning_rate": 4.779277084153737e-05, "logits/chosen": -0.5970318913459778, "logits/rejected": -0.4951511025428772, "logps/chosen": -299.33294677734375, "logps/rejected": -385.574462890625, "loss": 0.3647, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -8.652000427246094, "rewards/margins": 3.701908588409424, "rewards/rejected": -12.353909492492676, "step": 285 }, { "epoch": 0.41, "learning_rate": 4.7715308430821864e-05, "logits/chosen": -0.5532945394515991, "logits/rejected": -0.470702588558197, "logps/chosen": -289.81219482421875, "logps/rejected": -367.1038513183594, "loss": 0.3418, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -8.14123821258545, "rewards/margins": 4.235669136047363, "rewards/rejected": -12.376907348632812, "step": 290 }, { "epoch": 0.42, "learning_rate": 4.765242323034498e-05, "logits/chosen": -0.5983850359916687, "logits/rejected": -0.5276827216148376, "logps/chosen": -291.8949890136719, "logps/rejected": -367.7808532714844, "loss": 0.2942, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -6.658424377441406, "rewards/margins": 4.356478691101074, "rewards/rejected": -11.01490306854248, "step": 295 }, { "epoch": 0.43, "learning_rate": 4.757267633420931e-05, "logits/chosen": -0.5903723239898682, "logits/rejected": -0.470381498336792, "logps/chosen": -245.4473876953125, "logps/rejected": -322.921630859375, "loss": 0.2653, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -5.816718578338623, "rewards/margins": 3.5717501640319824, "rewards/rejected": -9.388467788696289, "step": 300 }, { "epoch": 0.43, "eval_logits/chosen": -0.6423404216766357, "eval_logits/rejected": -0.5549700856208801, "eval_logps/chosen": -277.4593811035156, "eval_logps/rejected": -346.16680908203125, "eval_loss": 0.3175624907016754, "eval_rewards/accuracies": 0.8372747302055359, "eval_rewards/chosen": -6.353503227233887, "eval_rewards/margins": 3.2359619140625, "eval_rewards/rejected": -9.589465141296387, "eval_runtime": 619.6776, "eval_samples_per_second": 0.952, "eval_steps_per_second": 0.119, "step": 300 }, { "epoch": 0.44, "learning_rate": 4.7491666499598794e-05, "logits/chosen": -0.4752650856971741, "logits/rejected": -0.4291561245918274, "logps/chosen": -308.3399963378906, "logps/rejected": -351.7706604003906, "loss": 0.2566, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -6.376312255859375, "rewards/margins": 2.960242986679077, "rewards/rejected": -9.336555480957031, "step": 305 }, { "epoch": 0.44, "learning_rate": 4.7409398259003744e-05, "logits/chosen": -0.5722283124923706, "logits/rejected": -0.43486547470092773, "logps/chosen": -271.27618408203125, "logps/rejected": -350.6097412109375, "loss": 0.3905, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -6.20401668548584, "rewards/margins": 3.1277387142181396, "rewards/rejected": -9.331754684448242, "step": 310 }, { "epoch": 0.45, "learning_rate": 4.732587621532214e-05, "logits/chosen": -0.4831075668334961, "logits/rejected": -0.3779456317424774, "logps/chosen": -275.4000549316406, "logps/rejected": -348.45404052734375, "loss": 0.3821, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": -7.6802659034729, "rewards/margins": 2.4858384132385254, "rewards/rejected": -10.166104316711426, "step": 315 }, { "epoch": 0.46, "learning_rate": 4.724110504160208e-05, "logits/chosen": -0.42866721749305725, "logits/rejected": -0.39401277899742126, "logps/chosen": -322.8934020996094, "logps/rejected": -362.2698669433594, "loss": 0.3577, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -8.4872407913208, "rewards/margins": 2.9192988872528076, "rewards/rejected": -11.406539916992188, "step": 320 }, { "epoch": 0.46, "learning_rate": 4.715508948078037e-05, "logits/chosen": -0.43140649795532227, "logits/rejected": -0.382727712392807, "logps/chosen": -309.9305114746094, "logps/rejected": -364.6683044433594, "loss": 0.4905, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": -8.847053527832031, "rewards/margins": 2.7970926761627197, "rewards/rejected": -11.644145965576172, "step": 325 }, { "epoch": 0.47, "learning_rate": 4.706783434541708e-05, "logits/chosen": -0.43298617005348206, "logits/rejected": -0.35435712337493896, "logps/chosen": -274.5734558105469, "logps/rejected": -323.9297790527344, "loss": 0.3598, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -7.113492012023926, "rewards/margins": 2.833127737045288, "rewards/rejected": -9.946619033813477, "step": 330 }, { "epoch": 0.48, "learning_rate": 4.6979344517426345e-05, "logits/chosen": -0.44348543882369995, "logits/rejected": -0.3648318350315094, "logps/chosen": -311.8634338378906, "logps/rejected": -370.2818298339844, "loss": 0.3383, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -8.36948299407959, "rewards/margins": 2.6521434783935547, "rewards/rejected": -11.021625518798828, "step": 335 }, { "epoch": 0.49, "learning_rate": 4.6889624947803195e-05, "logits/chosen": -0.4924314618110657, "logits/rejected": -0.3927859663963318, "logps/chosen": -286.1222229003906, "logps/rejected": -348.00445556640625, "loss": 0.2847, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -6.249931812286377, "rewards/margins": 3.2729618549346924, "rewards/rejected": -9.522893905639648, "step": 340 }, { "epoch": 0.49, "learning_rate": 4.679868065634656e-05, "logits/chosen": -0.4742973744869232, "logits/rejected": -0.3609599769115448, "logps/chosen": -267.9313049316406, "logps/rejected": -338.3818359375, "loss": 0.2471, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": -6.949034690856934, "rewards/margins": 3.703294277191162, "rewards/rejected": -10.652329444885254, "step": 345 }, { "epoch": 0.5, "learning_rate": 4.6706516731378406e-05, "logits/chosen": -0.4882968068122864, "logits/rejected": -0.4227726459503174, "logps/chosen": -256.69964599609375, "logps/rejected": -336.7755432128906, "loss": 0.2902, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": -6.316679954528809, "rewards/margins": 3.5853703022003174, "rewards/rejected": -9.902050971984863, "step": 350 }, { "epoch": 0.51, "learning_rate": 4.661313832945904e-05, "logits/chosen": -0.45601844787597656, "logits/rejected": -0.35979634523391724, "logps/chosen": -297.2603759765625, "logps/rejected": -384.87481689453125, "loss": 0.2746, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -8.420430183410645, "rewards/margins": 4.463082313537598, "rewards/rejected": -12.883511543273926, "step": 355 }, { "epoch": 0.51, "learning_rate": 4.65185506750986e-05, "logits/chosen": -0.39386382699012756, "logits/rejected": -0.28367191553115845, "logps/chosen": -309.7892150878906, "logps/rejected": -404.3554382324219, "loss": 0.3581, "rewards/accuracies": 0.875, "rewards/chosen": -10.17359447479248, "rewards/margins": 4.7968430519104, "rewards/rejected": -14.970438003540039, "step": 360 }, { "epoch": 0.52, "learning_rate": 4.642275906046475e-05, "logits/chosen": -0.42610305547714233, "logits/rejected": -0.32776302099227905, "logps/chosen": -308.56304931640625, "logps/rejected": -394.921142578125, "loss": 0.3434, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -9.843785285949707, "rewards/margins": 4.186667442321777, "rewards/rejected": -14.030451774597168, "step": 365 }, { "epoch": 0.53, "learning_rate": 4.632576884508662e-05, "logits/chosen": -0.38670024275779724, "logits/rejected": -0.29175546765327454, "logps/chosen": -317.8299560546875, "logps/rejected": -382.1197814941406, "loss": 0.3862, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -10.289070129394531, "rewards/margins": 4.806598663330078, "rewards/rejected": -15.095669746398926, "step": 370 }, { "epoch": 0.54, "learning_rate": 4.622758545555485e-05, "logits/chosen": -0.38157421350479126, "logits/rejected": -0.26931703090667725, "logps/chosen": -292.02069091796875, "logps/rejected": -388.94720458984375, "loss": 0.3453, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -7.471795558929443, "rewards/margins": 4.848240852355957, "rewards/rejected": -12.320035934448242, "step": 375 }, { "epoch": 0.54, "learning_rate": 4.612821438521805e-05, "logits/chosen": -0.3492121398448944, "logits/rejected": -0.2813674807548523, "logps/chosen": -305.62554931640625, "logps/rejected": -356.89898681640625, "loss": 0.2206, "rewards/accuracies": 0.925000011920929, "rewards/chosen": -8.157720565795898, "rewards/margins": 3.738807201385498, "rewards/rejected": -11.896527290344238, "step": 380 }, { "epoch": 0.55, "learning_rate": 4.602766119387544e-05, "logits/chosen": -0.3653913736343384, "logits/rejected": -0.2597258687019348, "logps/chosen": -311.99371337890625, "logps/rejected": -382.8812255859375, "loss": 0.3092, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -8.370109558105469, "rewards/margins": 3.6346893310546875, "rewards/rejected": -12.004796981811523, "step": 385 }, { "epoch": 0.56, "learning_rate": 4.592593150746576e-05, "logits/chosen": -0.3710058033466339, "logits/rejected": -0.2727561891078949, "logps/chosen": -288.8029479980469, "logps/rejected": -361.223388671875, "loss": 0.3955, "rewards/accuracies": 0.800000011920929, "rewards/chosen": -7.695769309997559, "rewards/margins": 3.3667006492614746, "rewards/rejected": -11.062471389770508, "step": 390 }, { "epoch": 0.56, "learning_rate": 4.5823031017752485e-05, "logits/chosen": -0.3490789532661438, "logits/rejected": -0.2746933102607727, "logps/chosen": -281.90704345703125, "logps/rejected": -338.3580322265625, "loss": 0.328, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": -7.884057521820068, "rewards/margins": 2.856874942779541, "rewards/rejected": -10.740933418273926, "step": 395 }, { "epoch": 0.57, "learning_rate": 4.571896548200542e-05, "logits/chosen": -0.4303611218929291, "logits/rejected": -0.3262333273887634, "logps/chosen": -261.8278503417969, "logps/rejected": -337.40521240234375, "loss": 0.2372, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": -6.010989189147949, "rewards/margins": 3.444431781768799, "rewards/rejected": -9.45542049407959, "step": 400 }, { "epoch": 0.57, "eval_logits/chosen": -0.4868081510066986, "eval_logits/rejected": -0.3944084346294403, "eval_logps/chosen": -282.0663757324219, "eval_logps/rejected": -350.36456298828125, "eval_loss": 0.32044681906700134, "eval_rewards/accuracies": 0.8355855941772461, "eval_rewards/chosen": -6.814201831817627, "eval_rewards/margins": 3.195034980773926, "eval_rewards/rejected": -10.009236335754395, "eval_runtime": 635.268, "eval_samples_per_second": 0.929, "eval_steps_per_second": 0.116, "step": 400 }, { "epoch": 0.58, "learning_rate": 4.5613740722678525e-05, "logits/chosen": -0.39522626996040344, "logits/rejected": -0.3030610680580139, "logps/chosen": -260.17266845703125, "logps/rejected": -320.7784118652344, "loss": 0.388, "rewards/accuracies": 0.8125, "rewards/chosen": -7.52327823638916, "rewards/margins": 3.621577739715576, "rewards/rejected": -11.144853591918945, "step": 405 }, { "epoch": 0.59, "learning_rate": 4.550736262708418e-05, "logits/chosen": -0.4773409962654114, "logits/rejected": -0.3466936945915222, "logps/chosen": -277.6498107910156, "logps/rejected": -373.0572509765625, "loss": 0.2318, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -6.803755760192871, "rewards/margins": 3.537554979324341, "rewards/rejected": -10.34131145477295, "step": 410 }, { "epoch": 0.59, "learning_rate": 4.5399837147063825e-05, "logits/chosen": -0.47985219955444336, "logits/rejected": -0.38592132925987244, "logps/chosen": -246.52841186523438, "logps/rejected": -332.86248779296875, "loss": 0.3023, "rewards/accuracies": 0.875, "rewards/chosen": -6.49072265625, "rewards/margins": 3.5066089630126953, "rewards/rejected": -9.997331619262695, "step": 415 }, { "epoch": 0.6, "learning_rate": 4.529117029865488e-05, "logits/chosen": -0.4670870900154114, "logits/rejected": -0.36911553144454956, "logps/chosen": -260.58758544921875, "logps/rejected": -342.8556823730469, "loss": 0.3234, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -7.461057186126709, "rewards/margins": 4.485016822814941, "rewards/rejected": -11.946073532104492, "step": 420 }, { "epoch": 0.61, "learning_rate": 4.518136816175419e-05, "logits/chosen": -0.5157058835029602, "logits/rejected": -0.4533557891845703, "logps/chosen": -296.40618896484375, "logps/rejected": -343.72515869140625, "loss": 0.4552, "rewards/accuracies": 0.7749999761581421, "rewards/chosen": -8.75401496887207, "rewards/margins": 2.8250133991241455, "rewards/rejected": -11.57902717590332, "step": 425 }, { "epoch": 0.61, "learning_rate": 4.5070436879777865e-05, "logits/chosen": -0.5464299917221069, "logits/rejected": -0.4678220748901367, "logps/chosen": -253.18478393554688, "logps/rejected": -321.50714111328125, "loss": 0.3686, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -7.0356292724609375, "rewards/margins": 3.3276419639587402, "rewards/rejected": -10.36327075958252, "step": 430 }, { "epoch": 0.62, "learning_rate": 4.495838265931754e-05, "logits/chosen": -0.5636313557624817, "logits/rejected": -0.4402475357055664, "logps/chosen": -268.5367126464844, "logps/rejected": -361.71697998046875, "loss": 0.2249, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": -7.160154819488525, "rewards/margins": 3.917328357696533, "rewards/rejected": -11.077482223510742, "step": 435 }, { "epoch": 0.63, "learning_rate": 4.4845211769793116e-05, "logits/chosen": -0.5847368240356445, "logits/rejected": -0.4880955219268799, "logps/chosen": -264.53619384765625, "logps/rejected": -339.5098571777344, "loss": 0.2691, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -7.8903937339782715, "rewards/margins": 3.62628173828125, "rewards/rejected": -11.51667594909668, "step": 440 }, { "epoch": 0.64, "learning_rate": 4.473093054310199e-05, "logits/chosen": -0.5972886085510254, "logits/rejected": -0.5250357389450073, "logps/chosen": -300.294189453125, "logps/rejected": -370.1164855957031, "loss": 0.3729, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -7.556139945983887, "rewards/margins": 4.214568614959717, "rewards/rejected": -11.770709037780762, "step": 445 }, { "epoch": 0.64, "learning_rate": 4.46155453732648e-05, "logits/chosen": -0.5448333024978638, "logits/rejected": -0.4635641574859619, "logps/chosen": -265.34027099609375, "logps/rejected": -342.3074035644531, "loss": 0.3015, "rewards/accuracies": 0.875, "rewards/chosen": -7.357121467590332, "rewards/margins": 4.111447334289551, "rewards/rejected": -11.468568801879883, "step": 450 }, { "epoch": 0.65, "learning_rate": 4.449906271606766e-05, "logits/chosen": -0.5757399797439575, "logits/rejected": -0.5023797750473022, "logps/chosen": -283.24224853515625, "logps/rejected": -336.78485107421875, "loss": 0.309, "rewards/accuracies": 0.875, "rewards/chosen": -8.1708345413208, "rewards/margins": 3.2705204486846924, "rewards/rejected": -11.441353797912598, "step": 455 }, { "epoch": 0.66, "learning_rate": 4.438148908870095e-05, "logits/chosen": -0.560847282409668, "logits/rejected": -0.4687212407588959, "logps/chosen": -262.0433349609375, "logps/rejected": -343.6734313964844, "loss": 0.2833, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -9.314498901367188, "rewards/margins": 3.861955165863037, "rewards/rejected": -13.176454544067383, "step": 460 }, { "epoch": 0.66, "learning_rate": 4.426283106939474e-05, "logits/chosen": -0.5405054092407227, "logits/rejected": -0.4934763014316559, "logps/chosen": -344.86859130859375, "logps/rejected": -396.4712219238281, "loss": 0.3875, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -11.307413101196289, "rewards/margins": 4.146143436431885, "rewards/rejected": -15.453557014465332, "step": 465 }, { "epoch": 0.67, "learning_rate": 4.414309529705064e-05, "logits/chosen": -0.5560386776924133, "logits/rejected": -0.47137537598609924, "logps/chosen": -328.85467529296875, "logps/rejected": -410.0113830566406, "loss": 0.4106, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -10.061564445495605, "rewards/margins": 4.763692855834961, "rewards/rejected": -14.825258255004883, "step": 470 }, { "epoch": 0.68, "learning_rate": 4.402228847087047e-05, "logits/chosen": -0.6119273900985718, "logits/rejected": -0.5192709565162659, "logps/chosen": -285.03509521484375, "logps/rejected": -359.7369689941406, "loss": 0.3005, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -9.684968948364258, "rewards/margins": 4.337248802185059, "rewards/rejected": -14.022216796875, "step": 475 }, { "epoch": 0.69, "learning_rate": 4.390041734998132e-05, "logits/chosen": -0.6753469109535217, "logits/rejected": -0.5812332034111023, "logps/chosen": -275.0848083496094, "logps/rejected": -365.5320739746094, "loss": 0.2345, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": -7.314253807067871, "rewards/margins": 4.565116882324219, "rewards/rejected": -11.879369735717773, "step": 480 }, { "epoch": 0.69, "learning_rate": 4.3777488753057494e-05, "logits/chosen": -0.6396665573120117, "logits/rejected": -0.5838817358016968, "logps/chosen": -311.5262145996094, "logps/rejected": -385.2457580566406, "loss": 0.2286, "rewards/accuracies": 0.9125000238418579, "rewards/chosen": -8.526026725769043, "rewards/margins": 3.986698627471924, "rewards/rejected": -12.512724876403809, "step": 485 }, { "epoch": 0.7, "learning_rate": 4.365350955793892e-05, "logits/chosen": -0.6760267019271851, "logits/rejected": -0.620951771736145, "logps/chosen": -313.78411865234375, "logps/rejected": -380.7740478515625, "loss": 0.3314, "rewards/accuracies": 0.8125, "rewards/chosen": -9.628170013427734, "rewards/margins": 3.3307273387908936, "rewards/rejected": -12.958897590637207, "step": 490 }, { "epoch": 0.71, "learning_rate": 4.3528486701246376e-05, "logits/chosen": -0.6738308072090149, "logits/rejected": -0.5901960730552673, "logps/chosen": -285.67437744140625, "logps/rejected": -362.6622009277344, "loss": 0.3119, "rewards/accuracies": 0.8125, "rewards/chosen": -9.267045974731445, "rewards/margins": 3.928302049636841, "rewards/rejected": -13.195347785949707, "step": 495 }, { "epoch": 0.71, "learning_rate": 4.3402427177993366e-05, "logits/chosen": -0.6766126155853271, "logits/rejected": -0.6122361421585083, "logps/chosen": -292.34326171875, "logps/rejected": -355.0762634277344, "loss": 0.4221, "rewards/accuracies": 0.800000011920929, "rewards/chosen": -8.092927932739258, "rewards/margins": 4.039681911468506, "rewards/rejected": -12.132608413696289, "step": 500 }, { "epoch": 0.71, "eval_logits/chosen": -0.7289233803749084, "eval_logits/rejected": -0.6465120911598206, "eval_logps/chosen": -292.57330322265625, "eval_logps/rejected": -368.62078857421875, "eval_loss": 0.31609994173049927, "eval_rewards/accuracies": 0.8474099040031433, "eval_rewards/chosen": -7.864894390106201, "eval_rewards/margins": 3.9699699878692627, "eval_rewards/rejected": -11.83486557006836, "eval_runtime": 640.9294, "eval_samples_per_second": 0.921, "eval_steps_per_second": 0.115, "step": 500 }, { "epoch": 0.72, "learning_rate": 4.327533804119476e-05, "logits/chosen": -0.6126315593719482, "logits/rejected": -0.5508357286453247, "logps/chosen": -282.2762756347656, "logps/rejected": -364.5578308105469, "loss": 0.316, "rewards/accuracies": 0.9375, "rewards/chosen": -7.242485046386719, "rewards/margins": 3.816096782684326, "rewards/rejected": -11.05858039855957, "step": 505 }, { "epoch": 0.73, "learning_rate": 4.314722640147222e-05, "logits/chosen": -0.6599456071853638, "logits/rejected": -0.5939927697181702, "logps/chosen": -323.53668212890625, "logps/rejected": -396.10833740234375, "loss": 0.2941, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -8.693486213684082, "rewards/margins": 4.0846662521362305, "rewards/rejected": -12.778152465820312, "step": 510 }, { "epoch": 0.74, "learning_rate": 4.301809942665625e-05, "logits/chosen": -0.6404463052749634, "logits/rejected": -0.5707911849021912, "logps/chosen": -307.43768310546875, "logps/rejected": -379.2852478027344, "loss": 0.4079, "rewards/accuracies": 0.862500011920929, "rewards/chosen": -9.3197660446167, "rewards/margins": 3.961516857147217, "rewards/rejected": -13.281283378601074, "step": 515 }, { "epoch": 0.74, "learning_rate": 4.28879643413853e-05, "logits/chosen": -0.65186607837677, "logits/rejected": -0.565552830696106, "logps/chosen": -310.2057189941406, "logps/rejected": -402.9633483886719, "loss": 0.2945, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -10.662870407104492, "rewards/margins": 4.87128210067749, "rewards/rejected": -15.534152030944824, "step": 520 }, { "epoch": 0.75, "learning_rate": 4.2756828426701426e-05, "logits/chosen": -0.648743212223053, "logits/rejected": -0.5434762239456177, "logps/chosen": -288.9256591796875, "logps/rejected": -386.8144836425781, "loss": 0.2866, "rewards/accuracies": 0.824999988079071, "rewards/chosen": -10.066654205322266, "rewards/margins": 5.4559478759765625, "rewards/rejected": -15.522601127624512, "step": 525 }, { "epoch": 0.76, "learning_rate": 4.2624699019643e-05, "logits/chosen": -0.6845074892044067, "logits/rejected": -0.6216086149215698, "logps/chosen": -298.6322937011719, "logps/rejected": -368.7320861816406, "loss": 0.3515, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -8.717851638793945, "rewards/margins": 4.482119560241699, "rewards/rejected": -13.199971199035645, "step": 530 }, { "epoch": 0.76, "learning_rate": 4.249158351283414e-05, "logits/chosen": -0.6760630011558533, "logits/rejected": -0.5973928570747375, "logps/chosen": -334.3371887207031, "logps/rejected": -412.73712158203125, "loss": 0.2374, "rewards/accuracies": 0.875, "rewards/chosen": -11.443408012390137, "rewards/margins": 5.742681503295898, "rewards/rejected": -17.18608856201172, "step": 535 }, { "epoch": 0.77, "learning_rate": 4.235748935407115e-05, "logits/chosen": -0.6781556010246277, "logits/rejected": -0.6254957914352417, "logps/chosen": -314.7560729980469, "logps/rejected": -381.03436279296875, "loss": 0.571, "rewards/accuracies": 0.7875000238418579, "rewards/chosen": -10.918689727783203, "rewards/margins": 4.366137981414795, "rewards/rejected": -15.284828186035156, "step": 540 }, { "epoch": 0.78, "learning_rate": 4.2222424045905774e-05, "logits/chosen": -0.703301727771759, "logits/rejected": -0.6157757639884949, "logps/chosen": -298.96417236328125, "logps/rejected": -374.37408447265625, "loss": 0.1956, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": -10.03012752532959, "rewards/margins": 4.996109962463379, "rewards/rejected": -15.026239395141602, "step": 545 }, { "epoch": 0.79, "learning_rate": 4.2086395145225456e-05, "logits/chosen": -0.6199958920478821, "logits/rejected": -0.5463480353355408, "logps/chosen": -297.76715087890625, "logps/rejected": -383.3893737792969, "loss": 0.3658, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -9.150928497314453, "rewards/margins": 4.323057651519775, "rewards/rejected": -13.473986625671387, "step": 550 }, { "epoch": 0.79, "learning_rate": 4.1949410262830525e-05, "logits/chosen": -0.6586548686027527, "logits/rejected": -0.598033607006073, "logps/chosen": -298.9888610839844, "logps/rejected": -358.97113037109375, "loss": 0.3435, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -7.790317535400391, "rewards/margins": 3.196791172027588, "rewards/rejected": -10.987109184265137, "step": 555 }, { "epoch": 0.8, "learning_rate": 4.1811477063008357e-05, "logits/chosen": -0.647259533405304, "logits/rejected": -0.5749470591545105, "logps/chosen": -283.01129150390625, "logps/rejected": -346.770263671875, "loss": 0.1992, "rewards/accuracies": 0.9125000238418579, "rewards/chosen": -8.713510513305664, "rewards/margins": 3.8825290203094482, "rewards/rejected": -12.596040725708008, "step": 560 }, { "epoch": 0.81, "learning_rate": 4.167260326310459e-05, "logits/chosen": -0.655887246131897, "logits/rejected": -0.5856226682662964, "logps/chosen": -353.996337890625, "logps/rejected": -423.36199951171875, "loss": 0.3649, "rewards/accuracies": 0.8125, "rewards/chosen": -7.39592981338501, "rewards/margins": 4.420022010803223, "rewards/rejected": -11.81595230102539, "step": 565 }, { "epoch": 0.81, "learning_rate": 4.1532796633091296e-05, "logits/chosen": -0.6630634069442749, "logits/rejected": -0.5999043583869934, "logps/chosen": -311.5198669433594, "logps/rejected": -382.8631896972656, "loss": 0.2865, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -7.376175880432129, "rewards/margins": 4.25052547454834, "rewards/rejected": -11.626702308654785, "step": 570 }, { "epoch": 0.82, "learning_rate": 4.139206499513231e-05, "logits/chosen": -0.6371768712997437, "logits/rejected": -0.552124559879303, "logps/chosen": -315.57696533203125, "logps/rejected": -386.8127746582031, "loss": 0.3623, "rewards/accuracies": 0.8374999761581421, "rewards/chosen": -7.146079063415527, "rewards/margins": 3.366482973098755, "rewards/rejected": -10.51256275177002, "step": 575 }, { "epoch": 0.83, "learning_rate": 4.1250416223145515e-05, "logits/chosen": -0.6766015887260437, "logits/rejected": -0.5857943296432495, "logps/chosen": -259.1465759277344, "logps/rejected": -349.38995361328125, "loss": 0.3386, "rewards/accuracies": 0.8125, "rewards/chosen": -6.320498466491699, "rewards/margins": 3.3259353637695312, "rewards/rejected": -9.64643383026123, "step": 580 }, { "epoch": 0.84, "learning_rate": 4.110785824236236e-05, "logits/chosen": -0.6103043556213379, "logits/rejected": -0.5319772958755493, "logps/chosen": -256.13446044921875, "logps/rejected": -331.8528137207031, "loss": 0.1824, "rewards/accuracies": 0.887499988079071, "rewards/chosen": -6.32611608505249, "rewards/margins": 3.831328868865967, "rewards/rejected": -10.157445907592773, "step": 585 }, { "epoch": 0.84, "learning_rate": 4.0964399028884394e-05, "logits/chosen": -0.6808601021766663, "logits/rejected": -0.5574377179145813, "logps/chosen": -255.42105102539062, "logps/rejected": -355.5065002441406, "loss": 0.2103, "rewards/accuracies": 0.8999999761581421, "rewards/chosen": -6.239567756652832, "rewards/margins": 4.258026599884033, "rewards/rejected": -10.497593879699707, "step": 590 }, { "epoch": 0.85, "learning_rate": 4.082004660923703e-05, "logits/chosen": -0.6791261434555054, "logits/rejected": -0.5921913385391235, "logps/chosen": -267.8202819824219, "logps/rejected": -350.561767578125, "loss": 0.2129, "rewards/accuracies": 0.9125000238418579, "rewards/chosen": -7.478656768798828, "rewards/margins": 4.9322614669799805, "rewards/rejected": -12.410919189453125, "step": 595 }, { "epoch": 0.86, "learning_rate": 4.067480905992045e-05, "logits/chosen": -0.7032105326652527, "logits/rejected": -0.600141167640686, "logps/chosen": -297.6030578613281, "logps/rejected": -384.7508850097656, "loss": 0.2602, "rewards/accuracies": 0.8500000238418579, "rewards/chosen": -9.610403060913086, "rewards/margins": 5.1422200202941895, "rewards/rejected": -14.7526216506958, "step": 600 }, { "epoch": 0.86, "eval_logits/chosen": -0.7646234631538391, "eval_logits/rejected": -0.6766347289085388, "eval_logps/chosen": -310.1993103027344, "eval_logps/rejected": -401.2223205566406, "eval_loss": 0.3386208117008209, "eval_rewards/accuracies": 0.8237612247467041, "eval_rewards/chosen": -9.627497673034668, "eval_rewards/margins": 5.467515468597412, "eval_rewards/rejected": -15.095013618469238, "eval_runtime": 635.6148, "eval_samples_per_second": 0.928, "eval_steps_per_second": 0.116, "step": 600 } ], "max_steps": 2100, "num_train_epochs": 3, "total_flos": 6.092901381557453e+17, "trial_name": null, "trial_params": null }